Edinburgh Research Explorer

Farming, foreign holidays, and vitamin D in Orkney

Citation for published version:Weiss, E, Zgaga, L, Read, S, Wild, S, Dunlop, M, Campbell, H, McQuillan, R & Wilson, J 2016, 'Farming,foreign holidays, and vitamin D in Orkney', PLoS ONE. https://doi.org/10.1371/journal.pone.0155633

Digital Object Identifier (DOI):10.1371/journal.pone.0155633

Link:Link to publication record in Edinburgh Research Explorer

Document Version:Publisher's PDF, also known as Version of record

Published In:PLoS ONE

Publisher Rights Statement:Copyright: © 2016 Weiss et al. This is an open access article distributed under the terms of the CreativeCommons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium,provided the original author and source are credited.

General rightsCopyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s)and / or other copyright owners and it is a condition of accessing these publications that users recognise andabide by the legal requirements associated with these rights.

Take down policyThe University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorercontent complies with UK legislation. If you believe that the public display of this file breaches copyright pleasecontact openaccess@ed.ac.uk providing details, and we will remove access to the work immediately andinvestigate your claim.

Download date: 22. Aug. 2021

RESEARCH ARTICLE

Farming, Foreign Holidays, and Vitamin D inOrkneyEmily Weiss1, Lina Zgaga2, Stephanie Read1, SarahWild1, Malcolm G. Dunlop3,Harry Campbell1, Ruth McQuillan1☯, James F. Wilson1,3☯*

1 Usher Institute for Population Health Sciences and Informatics, University of Edinburgh, Edinburgh,Scotland, 2 Public Health and Primary Care, Trinity College Centre for Health Sciences, Dublin, Ireland,3 MRCHuman Genetics Unit, MRC Institute of Genetics and Molecular Medicine, University of Edinburgh,Edinburgh, Scotland

☯ These authors contributed equally to this work.* jim.wilson@ed.ac.uk

AbstractOrkney, north of mainland Scotland, has the world’s highest prevalence of multiple sclerosis

(MS); vitamin D deficiency, a marker of low UV exposure, is also common in Scotland.

Strong associations have been identified between vitamin D deficiency and MS, and

between UV exposure and MS independent of vitamin D, although causal relationships

remain to be confirmed. We aimed to compare plasma 25-hydroxyvitamin D levels in Ork-

ney and mainland Scotland, and establish the determinants of vitamin D status in Orkney.

We compared mean vitamin D and prevalence of deficiency in cross-sectional study data

from participants in the Orkney Complex Disease Study (ORCADES) and controls in the

Scottish Colorectal Cancer Study (SOCCS). We used multivariable regression to identify

factors associated with vitamin D levels in Orkney. Mean (standard deviation) vitamin D

was significantly higher among ORCADES than SOCCS participants (35.3 (18.0) and 31.7

(21.2), respectively). Prevalence of severe vitamin D deficiency was lower in ORCADES

than SOCCS participants (6.6% to 16.2% p = 1.1 x 10−15). Older age, farming occupations

and foreign holidays were significantly associated with higher vitamin D in Orkney. Although

mean vitamin D levels are higher in Orkney than mainland Scotland, this masks variation

within the Orkney population which may influence MS risk.

IntroductionMultiple sclerosis is a chronic, complex disease with genetic, environmental and behaviouralfactors implicated in its aetiology [1]. Greater distance from the equator is associated withincreasing MS prevalence [2]; increasing latitude is also noted for weaker ultraviolet B (UVB)radiation which inhibits cutaneous production of vitamin D [3]. As such, one environmentalrisk factor is thought to be vitamin D deficiency, however, vitamin D is also a marker for expo-sure to UV radiation, the benefits of which may be independent of vitamin D production [4–6]. A variety of factors hinder or prevent UVB from reaching the earth's surface, including lati-tude and weather [7], or from initiating cutaneous vitamin D synthesis, such as sun protectioncreams and clothing cover.

PLOSONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 1 / 18

a11111

OPEN ACCESS

Citation:Weiss E, Zgaga L, Read S, Wild S, DunlopMG, Campbell H, et al. (2016) Farming, ForeignHolidays, and Vitamin D in Orkney. PLoS ONE 11(5):e0155633. doi:10.1371/journal.pone.0155633

Editor: Andrzej T Slominski, University of Alabama atBirmingham, UNITED STATES

Received: January 29, 2016

Accepted: May 1, 2016

Published: May 17, 2016

Copyright: © 2016 Weiss et al. This is an openaccess article distributed under the terms of theCreative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in anymedium, provided the original author and source arecredited.

Data Availability Statement: The data containindividual-level phenotypic information from isolatedhuman populations. The participants' signed consentand ethical permission do not allow that thisinformation be publicly available, as the identity ofspecific individuals can be inferred from such data.Requests for the data should be made to the QTL-executive at the Institute of Genetics and MolecularMedicine (IGMM) by contacting James F Wilson(mailto:jim.wilson@ed.ac.uk).

Funding: The SOCCS study was funded by aCancer Research UK Programme Grant (C348/A12076) (MGD, LZ, HC) (URL: http://www.cancerresearchuk.org). ORCADES was supported by

Although a recent systematic review of vitamin D status worldwide found that vitamin Dconcentrations do not appear to be dependent upon latitude [8], exposure to ultraviolet B(UVB) radiation from sunshine is the most potent source of vitamin D for humans [9]. In theUnited States, a latitudinal relationship exists between wintertime vitamin D and wintertimeUV doses [10], likely resulting from the few days in which vitamin D can be produced at suchlatitudes [3, 10]. This latitudinal relationship further reflects the latitudinal MS prevalence dis-tribution in the US [10]. Additionally, a significant relationship between regional UVB radia-tion and MS prevalence has been noted in France in the French farming population and theirfamilies [11]. Increasing MS prevalence was associated with decreasing ambient UVB; thetrend was additionally found to be stronger in both wintertime and in women [11]. A similarrelationship between decreasing UV and increasing prevalence has been identified in New-foundland [12] and Australia [13].

Controversy remains regarding the role of vitamin D in chronic conditions; whilst defi-ciency may have a causal role in the aetiology of some diseases it may also simply be a bio-marker for ill health. A body of evidence is however accumulating, suggesting a causal role forvitamin D in MS risk, pathology and progression [14–16]. A recent Mendelian randomisationstudy found genetically lowered 25-hydroxyvitamin D (25(OH)D) level by one standard devia-tion in log-transformed 25(OH)D was associated with a two-fold increased risk of MS [16].Interactions between vitamin D and the major genetic risk factor, HLA-DRB1�1501, have beenidentified [17], and several rare variants conferring MS risk have been found in the CYP27B1gene, which encodes an enzyme which catalyses the conversion of 25(OH)D to the bioactiveform [18]. Further, early exposure to vitamin D in-utero and in neonatal mice led to optimalnumbers of invariant natural killer T cells [19], deficiency of which are observed in MS patients[19, 20]. Alongside the month-of-birth effect, where children born after a winter gestation areat higher risk of MS [21], the role of adequate in-utero vitamin D increasingly appears to becritical for future autoimmunity.

The beneficial role of UV exposure independent of vitamin D production has been sup-ported in animal studies, using experimental autoimmune encephalomyelitis (EAE) as themodel for MS. Continuous treatment with UVB was found to suppress clinical signs of EAEwhich, although leading to slight elevations in serum 25(OH)D3, were insufficient to cause dis-ease suppression by vitamin D [4]. Furthermore, suppression of EAE was found to occur uponirradiation of mice to narrow-band UV light, with a wavelength of between 300 and 315 nmand a peak of effectiveness at 311 nm. As vitamin D requires a wavelength between 270 and300 nm, optimally between 295 and 300 nm to initiate cutaneous synthesis, the narrow bandUV supressing EAE had no effect on 25(OH)D levels [5], strongly suggesting a role of UVexposure independent of vitamin D. In MS, an Australian multi-centre case-control studyfound higher sun exposure and higher vitamin D levels to be independently associated withlower risk for first demyelinating events [6].

Scotland, between latitudes 54° and 60° north, has inadequate strength of sunshine betweenOctober and March for vitamin D synthesis [22]; a cloudy climate year-round further leads towidespread vitamin D deficiency [23] strongly indicating limited UV exposure in the popula-tion. The protective effect of supplementation and sunny holidays on 25(OH)D in Aberdeen, aScottish city at 57° north, has previously been noted in a study of postmenopausal women [24].Orkney, an isolated archipelago ten to sixty miles from the north coast of Scotland, is an areaof exceptionally high MS prevalence [25]. Seventeen of the 70 islands are inhabited with a pre-dominantly rural population totalling 21,349 at the 2011 census. The 2011 census also revealedan ongoing agricultural tradition with 10% of the workforce employed in agriculture or fishing.

As an independent risk factor, or as a marker of UV exposure, it is important to understandthe determinants of plasma vitamin D in the context of MS and other diseases of public health

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 2 / 18

the Chief Scientist Office of the Scottish Government(CZB/4/276, CZB/4/710) (JFW, HC) (URL: http://www.cso.scot.nhs.uk/). The Shetland and Orkney MultipleSclerosis Research Project supports a studentshipfrom which this work has resulted (EW under JFW).The funders had no role in the study design, datacollection and analysis, decision to publish, orpreparation of the manuscript.

Competing Interests: The authors have declaredthat no competing interests exist.

importance. In this study we aimed to describe vitamin D levels in Orkney (Fig 1). Thisinvolved identifying the prevalence of vitamin D deficiency in Orkney compared to the Scottishmainland, and establishing the determinants of circulating plasma 25-hydroxyvitamin D (25(OH)D) in Orkney.

Methods and Materials

Study PopulationsThe study population comprised Orkney Complex Disease Study (ORCADES) participants,recruited from 2005 to 2011, and Scottish Colorectal Cancer Study (SOCCS) controls, whichran from 1999 to 2006. Both these studies have been described in detail elsewhere [26, 27].Briefly, ORCADES, a cross-sectional genetic epidemiology study concerned with identifyinggenetic factors that influence complex disease, comprised 2078 participants with at least oneOrcadian grandparent. SOCCS, a case-control study of colorectal cancer in Scotland, included2235 adult controls without colorectal cancer, identified from the Community Health Index inScotland as being aged within 5 years their matched case, of the same sex and living in thesame area. All participants provided written, informed consent prior to participation. Bothstudies have ethical approval from NHS Orkney, NHS Grampian or NHS Lothian.

25-hydroxyvitamin D measurementFasting blood samples were drawn from ORCADES participants using the Sarstedt Monovettesystem. Samples were processed and transferred for storage at -400 C, and later -800 C, untilanalysis. Blood was drawn from all SOCCS participants, processed and transferred for storageat -800 C. Both studies ran over multiple years, and therefore measures per month compriseblood drawn in multiple Januarys, multiple Februarys and so on.

Vitamin D status is determined by measuring circulating 25(OH)D which is generally con-sidered the best indicator of vitamin D status [28]. Both 25(OH)D2 and 25(OH)D3 were mea-sured using the liquid chromatography-tandem mass spectrometry (LC-MS/MS) method. Thetotal of the two measures was taken for total circulating 25(OH)D, however most samples from

Fig 1. Map of Orkney in relation to the Scottish mainland and north-west periphery of Europe, with57th and 59th degrees of latitude.

doi:10.1371/journal.pone.0155633.g001

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 3 / 18

both studies contained no 25(OH)D2. The lower limit of detection using LC-MS/MS was 10nmol/L. All samples were measured in the same laboratory following standard protocols; qual-ity control procedures were performed according to current best evidence for 25(OH)D mea-surement in population studies [29].

A range of cut-offs to define sufficiency and deficiency have been proposed, however in linewith other recent studies we considered circulating 25(OH)D of 50 nmol/L or over to be suffi-cient [30], 25 to 50 nmol/L to be at risk of deficiency or insufficient, and deficiency to be lessthan 25 nmol/L [31]. We additionally explored those at the lower end of deficiency, where weconsidered circulating 25(OH)D below 12.5 nmol/L to be severely deficient [32].

Lifestyle factorsORCADES participants attended clinics where several biometric measures were recorded. Eachparticipant also completed a medical and lifestyle questionnaire from which vitamin D intake,physical activity (PA) and socioeconomic status (SES) were derived.

BMI was calculated as kg/m2 and treated as a continuous variable. Height was taken withoutshoes and weight wearing only light clothing. Based on questionnaire data we derived a variableencompassing work and leisure time PA throughout the year. Participants classified leisureactivity as either 1) light (mostly sitting, light housework) or 2) moderate exercise; and likewisework activity as 1) mostly sitting, 2) mostly standing, 3) manual work or 4) heavy manualwork. Dietary vitamin D intake was estimated from two self-administered food frequency ques-tionnaires, the cardiovascular disease questionnaire (CVDQ) and bone density questionnaire(BDQ). The CVDQ was treated as the primary source due to the higher response rate, howeverinformation contained in the BDQ that was not present in the CVDQ was merged to create themost comprehensive variable possible. Further, a research nurse-administered drug question-naire sought information about medications and dietary supplements. Participants alsodescribed their frequency of taking holidays within or outside the UK (never, less than once ayear, once a year, more than once a year). Although the Scottish Index for Multiple Deprivation(SIMD) is available for Orkney, the scattered and heterogeneous population means that con-centrations of poverty or affluence are difficult to identify; moreover neighbouring islands aregrouped together in units thus there is little discrimination [33]. Principal Components Analy-sis (PCA) is a statistical technique to reduce a number of variables into a few independentdimensions reflecting the underlying patterns in the data, and was used here to construct SESindices [34]. To establish a variable that differentiates between individuals, three SES variableswere thereby derived from 10 questionnaire items with significant loadings in PCA (S1 Table).Additionally, we applied an occupational prestige score to questionnaire occupation informa-tion which was then included in the PCA [35]. Holidays, car age and council tax band loadedsignificantly onto the first component; housing tenure, length of car ownership and highestqualification loaded significantly onto the second component, and job prestige score, years ineducation and supervisory role at work loaded significantly onto the third. This third compo-nent captures “non-traditional” lifestyles reflecting managerial, administrative and professionalpositions in contrast to traditional agricultural work. Time outside in summer was summedfrom participants’ estimates of average time spent without a roof covering on summer workand leisure days. Data from SOCCS included age, sex, month of blood sample and 25(OH)Dmeasurement.

Statistical analysisWematched the ORCADES and SOCCS datasets on age to within two years (Table 1) toremove any differences arising from age structures. Matching was carried out blind excepting

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 4 / 18

dataset of origin and age. Because of the large effect of season on vitamin D levels, we standard-ised 25(OH)D measurements to the month of May; values obtained thereby represent thosethat would be expected if every sample were drawn in May [32]. The mean of monthly meansin the mainland Scottish data is 34.4 nmol/L; in the Orcadian data the mean of monthly meansis 37.7 nmol/L. The May means are 33.8 nmol/L and 35.5 nmol/L, respectively. We used May-standardised measurements for all analyses concerned with determinants of vitamin D, andalso to compare Orkney and mainland Scotland in deficiency levels. For analyses concernedwith vitamin D and time of year we used crude 25(OH)D measures. Data are presented asmean (standard deviation).

We plotted crude 25(OH)D by location as a density plot and by month, and comparedusing a t-test. We compared vitamin D by age group using t-tests. To compare levels of defi-ciency in Orkney and mainland Scotland, we divided participants into groups of deficiency andplotted May-adjusted vitamin D for each deficiency group and location and tested for differ-ences using chi-square tests.

For determinants of May-adjusted 25(OH)D in Orkney, we ran a series of bivariable modelsof May-adjusted 25(OH)D against environmental and demographic variables of interest.Those that were significant were put into a multivariable model. These significant variablescomprised BMI, age at venepuncture, foreign holidays, PA, SES, dietary vitamin D and workingstatus. Sex was also a covariate. A large percentage of missing data (S2 Table) was imputedusing Multiple Imputation of Chained Equations (MICE)[36] after excluding 28 individualswith missing outcome data. We ran 68 cycles of 100 imputations and pooled the results in a lin-ear regression model. We ran the same model using complete cases only. Statistical tests weretwo-sided with p<0.05 taken as significant. Finally, we applied a one-way ANOVA to comparemean May-adjusted 25(OH)D across the three groups of participants that we identified as aresult of our analyses.

We assessed homoscedasticity by inspection of a QQ plot, and distribution of residualsusing a histogram with superimposed normal curve. Independence was checked using the Dur-bin Watson statistic, multicollinearity and outliers using the vif statistic and Cook’s distance,respectively. All analyses were conducted using R software version 3.2.0 [37].

ResultsFor this study, 64 individuals were excluded from ORCADES who were not resident in Orkney,10 who had MS, as well as 8 duplicate measures. Characteristics of ORCADES participants arepresented in Table 2. Twenty-three people were excluded from the Scottish Colorectal Cancer

Table 1. Distribution of age and crude vitamin D in age-matched Orkney andmainland Scotland datasets. The mainland dataset excludes peoplefrom above the 57th degree of latitude.

Orkney Mean 25OHD Mainland Mean 25OHD

No (%) (nmol/L) No (%) (nmol/L)

Participants 1453 36.2 1453 35.4

< 40 46 (3.17) 26.8 46 (3.17) 36.5

40–49 263 (18.1) 33.4 263 (18.1) 39.5

50–59 399 (27.5) 35.7 400 (27.5) 38.7

60–69 466 (32.1) 37.9 464 (31.9) 33.3

70 + 279 (19.2) 39.5 280 (19.3) 30.4

Sex (male) 590 (40.6) 37.2 794 (54.6) 35.4

Sex (female) 863 (59.4) 35.6 659 (45.4) 35.4

doi:10.1371/journal.pone.0155633.t001

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 5 / 18

Study who lived above the 57th degree of latitude. For comparison analyses, data were age-matched giving a final count of 1453 people in each dataset.

Using age-matched data we compared mean 25(OH)D in Orkney to mainland Scotland(Fig 2). Orkney had significantly higher crude 25(OH)D than mainland Scotland (Orkney35.3 (18.01), Mainland 31.7 (21.18), t(2800) = -4. 93, p = 8.5 x 10−7). Mean 25(OH)D washigher in Orkney for every month except August when the mainland peaks at ~50 nmol/L.The distribution of vitamin D levels is shifted to the right in Orkney (Fig 3). We comparedvitamin D in Orkney and mainland Scotland by age group (Table 3). Results for each agegroup are significantly different, however in the under 40s, 40 to 49 and 50 to 59 age groups,mainland Scotland has higher vitamin D, whilst in the 60 to 69 and over 70s age groups, Ork-ney has significantly higher vitamin D. Comparing Orkney with the mainland by deficiencygroup, we found that more people in Orkney had insufficient vitamin D, χ2 (1, N = 2863) =30.3, p = 3.8 x 10−8; however, Orkney had significantly fewer people with circulating 25(OH)D of<12.5nmol/L (severely deficient) compared with the mainland, χ2 (1, N = 2863) = 64.3,p = 1.1 x 10−15 (Fig 4).

To explore correlates of vitamin D in Orkney we ran two multivariable regression analyses,using both imputed data and complete cases. Each model yielded similar results (Table 4). Var-iables significantly associated with higher 25(OH)D included lower BMI, more foreign holi-days, older age and increased PA. Associated with lower 25(OH)D was the “non-traditional”SES grouping.

Table 2. Characteristics of ORCADES Study participants (n = 1972).

Characteristic No or mean SD % or range

Age at venepuncture (years) 53.4 15.3 16.5–100.2

Sex

Female 1191 - 60.4

Male 781 - 39.6

Body Mass Index (kg/m2) 27.7 4.9 16.9–53.9

Vitamin D intake (μg) 4.4 3.1 0.00–34.1

Physical activity* 5.1 1.2 3.0–8.0

Summer minutes outdoors 223 142 4.8–900

Working 1367 - 69.3

Retired 547 - 27.7

Holidays outside the UK

< once a year 1472 - 74.6

Once a year 329 - 16.7

> once a year 105 - 5.3

Years in education 16 1.2 1.0–23

Qualification level

O & standard grades, CSE 275 - 13.9

Highers, A levels 787 - 39.9

Certificates/diplomas 739 - 37.5

Bachelor/Master degree 88 - 4.5

Doctorate 13 - 0.7

* Physical activity scored from 1 (mostly sitting; inactive) to 4 (heavy manual labour; active) across different

domains within work and leisure. Each score is the sum of answers creating an individual value for each

participant.

doi:10.1371/journal.pone.0155633.t002

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 6 / 18

The association between older age and higher vitamin D required further exploration; webegan by comparing foreign holiday-takers and their non-holidaying counterparts. We foundthat people over 50 were significantly more likely to take foreign holidays at least once a yearcompared to people under 50 (Table 5, Fig 5) (χ2(1) = 6.4, p = 0.0083). We termed this the‘Saga’ effect. Additionally, we found that foreign holidays had a stronger effect on people over50 who had their blood drawn in the low vitamin D (weaker UVB) season (October to March)

Fig 2. Mean crude vitamin D concentration (nmol/L) per month by location, using age-matched data.Orkney’s mean vitamin Dis higher than the mainland for every month except August and November. Each study ran over consecutive years and measurementstaken in the samemonth each year were pooled.

doi:10.1371/journal.pone.0155633.g002

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 7 / 18

compared with people who had their blood drawn in the high vitamin D (stronger UVB) sea-son (April to September) (Table 6).

To explore under-40s further, we ran the same analyses comparing those who do and donot holiday outside the UK that were used for over-50s (S3 Table). The same results were sig-nificant, excepting body fat percentage, socio-economic status 2, age, and summer minutesspent outside.

Fig 3. Comparison of May-adjusted vitamin D distribution in Orkney andmainland Scotland usingage-matched data. The distribution for Orkney is to the right of the distribution for the mainland, reflecting thelower prevalence of severe deficiency, and peaks higher.

doi:10.1371/journal.pone.0155633.g003

Table 3. Comparison of vitamin D in Orkney andmainland Scotland by age group.

N = Orkney; Scotland Orkney mean crude 25(OH)D Mainland mean crude 25(OH)D t-test p-value

< 40 46; 46 26.8 36.5 -2.67 0.009

40–49 263; 263 33.4 39.5 -3.13 0.002

50–59 399; 400 35.7 38.7 -1.97 0.049

60–69 466; 464 37.95 33.28 3.59 0.0004

70+ 279; 280 38.51 30.39 4.84 1.6x10-6

doi:10.1371/journal.pone.0155633.t003

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 8 / 18

We also identified a ‘farmer effect’ (Table 7). Participants employed in “traditional” agricul-tural occupations that kept them outdoors had significantly higher mean vitamin D levels thanparticipants in non-traditional professions that kept them indoors (farmers 36.9 (18.0), non-farmers 33.8 (11.8), t(383) = 2.46, p = 0.014). Further, farmers tended to be older.

To test for differences in mean vitamin D across the ‘Saga’ group, farmers, and non-farmerswho are under 50 and do not take foreign holidays, we did a one-way ANOVA. This ANOVA

Fig 4. Comparison of percentage of people in May-adjusted vitamin D deficiency groups by location. The main differencesoccur in the severely deficient group (<12.5 nmol/L) which has significantly fewer people from the Orkney sample (χ2(1) = 64.2,p = 1.10 x 10−15), and the ‘at risk’ category (25-<50 nmol/L) which has significantly fewer people from the mainland Scottish sample(χ2(1) = 30.3, p = 3.78 x 10−8).

doi:10.1371/journal.pone.0155633.g004

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 9 / 18

was significant: Welch's F(2, 481.99) = 54.49, p = 2.2 x 10−16, and we therefore concluded thatvitamin D varies significantly across these groups with people over 50 who take foreign holi-days having higher vitamin D than farmers, who had higher vitamin D than non-farmers andpeople under 50 who remain in the UK (Fig 6).

DiscussionWe aimed to compare vitamin D levels in Orkney and mainland Scotland, and to identify thedeterminants of Orkney vitamin D. Definitions of vitamin D deficiency are much discussed,however it has been proposed that circulating 25(OH)D above 50 nmol/L are sufficient [30,38]. Vitamin D status in Scotland has been previously explored [32]. Deficient and high riskindividuals comprised 63.4% in the previous study; in our Orkney dataset deficient and highrisk individuals comprised 65.3%. However, people with severe deficiency (<12.5 nmol/L)comprised 11.8% of the former study and only 5.0% of the latter. Therefore although perhapsinitially surprising that mean vitamin D was higher in Orkney despite the higher latitude, thesmaller percentage of people with severe deficiency in Orkney led to this elevation. In bothdatasets the majority are either deficient or at risk of deficiency which could have significanthealth implications [9].

Ability to synthesise vitamin D decreases with age [39]; it is well established that older age isassociated with lower vitamin D and increased deficiency risk [32, 40]. However, we found thatlifestyle factors particular to Orkney contributed to better vitamin D in older compared toyounger people.

Table 4. Results of linear regression for complete cases and imputed data using May-adjusted vitamin D as the outcome.

Multivariable models

Model 1a (n = 628) Model 2b (n = 1949)

Predictors Est (95% CI) p-value Est (95% CI) p-value

Intercept 32.4 (18.5, 46.2) 5.25x10-6 30.3 (22.2, 38.3) 2.2x10-13

Body mass index (kg/m2) -0.75 (-1.02, -0.47) 1.95x10-7 -0.54 (-0.70, -0.38) 7.5x10-11

Holidays outside the UK

< once a year -0.93 (-4.38, 2.52) 0.59 0.75 (-1.34, 2.85) 0.48

Once year 5.03 (0.20, 9.86) 0.041 6.47 (3.47, 9.47) 0.000024

> once a year 18.7 (11.3, 26.2) 1.04x10-6 13.5 (9.07, 18.0) 3.4x10-9

Age at venepuncture 0.24 (0.11, 0.36) 0.0003 0.14 (0.07, 0.22) 0.00030

Physical activity 1.66 (0.58, 2.75) 0.003 1.42 (0.65, 2.19) 0.00032

Socio-economic status 3 (“non-traditional”) -2.10 (-3.57, -0.64) 0.005 -1.74 (-2.71, -0.78) 0.00043

Summer minutes outside 0.0046 (-0.00398, 0.013) 0.29 0.006 (-0.00028, 0.012) 0.062

Socio-economic status 2 0.32 (-1.30, 1.94) 0.70 0.69 (-0.34, 1.72) 0.19

Vitamin D intake (μg) 0.201 (-0.18, 0.60) 0.30 0.14 (-0.17, 0.46) 0.37

Working (not retired) -2.07 (-6.61, 2.47) 0.37 -0.18 (-2.45, 2.09) 0.88

Sex (male) 1.06 (-1.55, 3.67) 0.43 -0.12 (-1.77, 1.52) 0.88

Socio-economic status 1 -0.013 (-1.85, 1.82) 0.99 -0.075 (-1.17, 1.02) 0.89

a Model 1 constructed using complete cases in the original dataset, R2 = 0.204.b Model 2 constructed using 68 datasets with missing data completed by imputation (100 cycles), R2 = 0.111 (People missing outcome data were

excluded from the imputation model).

Socio-economic status was derived from principal components analysis.

doi:10.1371/journal.pone.0155633.t004

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 10 / 18

Participants who took foreign holidays had higher vitamin D than those who did not takeforeign holidays; furthermore, taking foreign holidays increased by age group. Less than 20%of under-40s in our Orkney sample took foreign holidays, whereas over 30% in the 70 and overgroup reported leaving the UK at least once a year. Weak sunshine within the UK leads tofewer opportunities for UV exposure and UVB-mediated vitamin D synthesis, and the effect offoreign holiday sun exposure has been previously associated with improved vitamin D in Scot-land [24]. Orkney mean vitamin D remained higher than mainland Scotland throughout theyear with the exception of two months: August and November. Although we were unable toexplore this further, mainland Scotland’s August elevation of vitamin D could be attributed tothe effect of holidays, following one month after school holidays. In November, however, main-land Scotland’s mean vitamin D levels were only minimally higher than Orkney. We foundthat mean vitamin D in people 50 and over taking foreign holidays was significantly higherthan vitamin D levels in the rest of the sample. Foreign holidays contributed more to vitaminD levels on blood drawn in months of weaker UVB. We were unable to explore at what time ofyear people take holidays, however this finding suggests that foreign holidays become moreimportant as a source of UV exposure and therefore vitamin D for Orkney residents in winter.As older people tend to have more freedom to travel outside peak season, they are able to derivemost benefit by seeking sun in seasons of scarce sunshine in Orkney. People over 50 who takeforeign holidays were found to differ from those who do not take holidays mainly in variablesdenoting financial security. They were also more likely to have lower BMI and body fat per-centage suggesting possible healthier lifestyles than their non-holidaying counterparts.

Table 5. Comparison of people over 50 who holiday outside the UK at least once a year (n = 281) and people over 50 who holiday outside the UKless than once a year or never (n = 851). Unpaired t-tests applied to continuous data; chi-square tests applied to categorical data.

Over-50s, holiday Over-50s, no holiday t-test or p-value

No (%) or Mean (SD) No (%) or Mean (SD) Chi-square

Socio-economic status 1 1.01 (0.76) -0.30 (0.92) -22.6 <2.2x10-16

Job prestige score 0.32 (1.08) -0.19 (0.96) -7.09 5.1x10-12

Socio-economic status 3 (“non-traditional”) 0.18 (1.05) -0.29 (0.92) -6.50 2.1x10-10

Supervisory role at work

Yes 186 (66.4) 377 (45.1)

No 94 (33.6) 459 (54.9) 38.2 6.4x10-10

Years in education 15.9 (1.26) 15.5 (1.14) -5.39 1.1x10-6

Body mass index (kg/m2) 27.6 (3.91) 28.9 (5.03) 4.41 1.2x10-5

Highest qualification

O & standard grades, CSE* 33 (11.8) 102 (12.1)

Highers, A levels** 114 (40.7) 246 (29.3)

Certificates/diplomas 113 (40.4) 461 (54.8)

Bachelor/Master/PhD 20 (7.14) 32 (3.81) 22.2 5.8x10-5

Summer minutes 252 (146) 214 (136) -3.38 0.001

Age 62.9 (7.62) 64.4 (8.63) 2.55 0.01

Bodyfat % 33.7 (7.51) 34.9 (8.07) 2.34 0.02

Socio-economic status 2 0.43 (0.76) 0.32 (0.76) -2.06 0.04

Vitamin D intake (μg) 5.09 (3.60) 4.78 (3.13) -1.12 0.26

Physical activity 5.21 (1.20) 5.16 (1.26) -0.37 0.71

* School examinations taken in the UK ~16 years of age

** School examinations taken in the UK ~18 years of age

doi:10.1371/journal.pone.0155633.t005

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 11 / 18

Fig 5. Percentage of people per age group in ORCADES who holiday outside the UK at least once a year. People over 50 takesignificantly more holidays than those under 50.

doi:10.1371/journal.pone.0155633.g005

Table 6. Mean May-adjusted vitamin D (nmol/L) according to season of venepuncture (high season(April-September, n = 96) vs low season (October-March, n = 185)) in people over 50 who take a holi-day outside the UK at least once a year. Linear regression with May-adjusted vitamin D as the outcome.

Std Beta Beta (95% CI) Std error p-value

High season over-50s 0.17 7.86 (3.57–12.2) 2.18 0.00035

Low season over-50s 0.19 8.33 (5.23–11.4) 1.57 1.75x10-7

doi:10.1371/journal.pone.0155633.t006

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 12 / 18

We also examined under 40s, the age group in which MS is most likely to be diagnosed andpregnancies are most likely to occur, thereby potentially conferring risk to the unborn child. Inthis group, we found that the main differences in those who do or do not take holidays out ofthe UK are related to financial security. Only 75 of the 400 people who are under 40 reportedleaving the UK for a holiday at least once a year, therefore, in the most at-risk group, inade-quate UV exposure in Orkney is compounded by a low prevalence of foreign holidays.

The ‘non-traditional’ SES group derived from PCA, comprising job prestige score, educa-tion years and supervisory role at work, was associated with vitamin D. These variables, reflect-ing “non-traditional” lifestyles of managerial, administrative and professional occupations incontrast to traditional agricultural work, related to farmers and non-farmers. Farmers werefound to be slightly less educated, possibly as a result of leaving school at the minimum leavingage about half a year to a year before the first set of examinations. Farmers were also less likelyto describe themselves as having a supervisory role at work than non-farmers, and also had aslightly lower-than-average job prestige score. The inverse association between vitamin D anda higher score in this variable means that farmers, who scored lower, had better vitamin D.Farmers in our cohort were also more likely to be older than non-farmers, further contributingto our finding of vitamin D increasing with age.

Physical inactivity and obesity have previously been related to low vitamin D in a largeAmerican cohort [41]; the association between lower BMI and higher 25(OH)D is also wellestablished [42]. The mechanism for lower vitamin D in the presence of higher BMI is thoughtto be a result of increased deposition of vitamin D in body fat [43], making BMI a proxy foradiposity. However, BMI does not distinguish between body fat and fat free mass, and is not

Table 7. Comparison of farmers (n = 265) and non-farmers (n = 1649) on variables of interest in Orkney. Farmers are anyone who identified their pri-mary profession as farmer. Unpaired t-tests applied to continuous data; chi-square tests applied to categorical data.

Farmers Non-farmers t-test or p-value

No (%) or Mean (SD) No (%) or Mean (SD) Chi-square

Age 60.7 (11.4) 52.4 (11.5) -8.86 <2.2x10-16

Socio-economic status 3 (“non-traditional”) -0.30 (0.58) 0.10 (1.02) 8.77 <2.2x10-16

Years in education 15.4 (1.01) 16.1 (1.24) 10.16 <2.2x10-16

Socio-economic status 1 -0.44 (1.00) 0.09 (0.97) 7.86 5.4x10-14

Physical activity 5.87 (1.23) 5.10 (1.21) -7.60 8.9x10-13

Highest qualification

O & standard grades, CSE* 24 (9.2) 251 (15.3)

Highers, A levels** 84 (32.1) 702 (42.8)

Certificates/diplomas 153 (58.4) 586 (35.8)

Bachelor/Master/PhD 1 (0.4) 100 (6.1) 55.9 4.3x10-12

Bodyfat % 30.4 (8.7) 33.1 (8.6) 4.51 9.1x10-6

Supervisory role at work

Yes 109 (41.3) 866 (53.2)

No 155 (58.7) 763 (46.8) 12.8 0.0003

Socio-economic status 2 0.28 (0.92) 0.07 (0.92) -3.39 0.0008

Body mass index (kg/m2) 28.4 (4.80) 27.7 (4.96) -2.12 0.04

Vitamin D intake (μg) 4.71 (2.93) 4.40 (3.20) -1.29 0.19

Summer minutes 236 (151) 221 (141) -1.21 0.23

* School examinations taken in the UK at ~16 years of age

** School examinations taken in the UK at ~18 years of age

doi:10.1371/journal.pone.0155633.t007

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 13 / 18

always a reliable indicator of adiposity in people with lower body fat but greater muscle mass[44]. The farmers in our cohort reflected this difficulty: they had lower body fat percentage buthigher BMI than non-farmers. We found farmers were more active and leaner than non-farm-ers and it is perhaps therefore fair to assume they have higher than average muscle mass. Nev-ertheless, in the multivariable models BMI followed what is expected.

Farmers had mean vitamin D significantly higher than non-farmers, but significantly lowerthan people over 50 who take foreign holidays. Summer minutes outside was not significant inthe multivariable analyses, however we performed a t-test for farmers versus non farmers andtime spent outside. Farmers, perhaps unsurprisingly, were found to spend significantly moretime outside than non-farmers which enables maximisation of even the smallest window ofvitamin-D strength sunshine. Although Zgaga et al. found higher vitamin D consumption ledto slight improvements in plasma vitamin D [32], we found that diet was not associated withvitamin D in Orkney. However, difficulties involved in building a variable with the availabledata likely contributed to this finding.

Both studies were recruited on an ‘opt in’ basis which may result in the samples representinga healthier than average population; however a strength of this study was the large number of

Fig 6. Mean May-adjusted vitamin D (nmol/L) in different groups in ORCADES. 95% confidence intervalbars are given. “Saga” refers to people over 50 who take a holiday outside the UK at least once a year;“Farmer” to people who identified their primary profession as farming; “Other” is people under 50 who take aholiday outside the UK less than once a year, and are not farmers.

doi:10.1371/journal.pone.0155633.g006

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 14 / 18

participants in each cohort. Particularly novel was the number of farmers in our Orkneycohort, enabling us to explore vitamin D in a select group within a rural population which isnot often studied. All vitamin D samples from both cohorts were analysed in the same labora-tory using the same procedures, helping maintain consistency and reliability of results. We hadaccess to a variety of detailed measures to explore vitamin D in Orkney; however data on timespent outside and vitamin D intake were somewhat limited and these may thus be morestrongly implicated in vitamin D than we were able to detect. Nevertheless, we found signifi-cant effects and reliable relationships for vitamin D with foreign holidays, BMI, physical activ-ity and age.

ConclusionMean vitamin D in Orkney was higher than mainland Scotland, driven largely by a lower per-centage of individuals with severe deficiency in Orkney. Overall concentrations in both cohortswere low with most people either deficient or at risk of deficiency, suggesting that UV exposurefor much of the year is low. Older Orkney residents were more likely to have better vitamin Dthan younger residents, largely resulting from the ‘Saga’ and ‘Farmer’ effects. Those most atrisk of deficiency in Orkney were under 40, an age group traditionally considered at lower riskof deficiency, but at increased risk of MS diagnosis. Within these main child-bearing years, alack of UV exposure and vitamin D deficiency may result in significant autoimmune implica-tions for offspring. The significant contribution of foreign holidays to Orkney vitamin D isconsistent with the findings of previous UK studies; the importance of foreign holidays in pro-viding adequate UV exposure to UK residents is underappreciated. We have found that youn-ger ages are more at risk from inadequate UV exposure and vitamin D deficiency in Orkney, acounty with a very high prevalence of MS. Further research exploring the relationship betweenvitamin D and quantitatively-measured exposure to UV radiation from sunshine and physicalactivity, as well as more detailed dietary information in Shetland, the most northerly UKcounty with an MS prevalence lower than Orkney, would help further elucidate the roles of UVexposure and vitamin D as MS risk factors in these islands.

Supporting InformationS1 Table. Principal components for socio-economic status variables. The most significantloadings are in bold.(DOCX)

S2 Table. Missing data in the Orkney dataset in variables of interest(DOCX)

S3 Table. Comparison of people under 40 who holiday outside the UK at least once a year(n = 75) and people under 40 who holiday outside the UK less than once a year or never(n = 325). Unpaired t-tests applied to continuous data; chi-square tests applied to categoricaldata.(DOCX)

AcknowledgmentsWe thank Craig Nicol for his assistance with the figures; the ORCADES data collection andadministrative teams and the people of Orkney.

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 15 / 18

Author ContributionsConceived and designed the experiments: EW JFW RM. Performed the experiments: EW. Ana-lyzed the data: EW LZ SR. Contributed reagents/materials/analysis tools: HC MGD JFW.Wrote the paper: EW LZ SR SWHCMGD RM JFW.

References1. Dyment DA, Ebers GC, Dessa Sadovnick A. Genetics of multiple sclerosis. Lancet Neurol. 2004; 3

(2):104–10. PMID: 14747002

2. Simpson S Jr, Blizzard L., Otahal P, Van der Mei I, Taylor B. Latitude is significantly associated with theprevalence of multiple sclerosis: a meta-analysis. J Neurol Neurosurg Psychiatry. 2011.

3. Webb A, Kline L, Holick M. Influence of season and latitude on the cutaneous synthesis of vitamin D3:exposure to winter sunlight in Boston and Edmonton will not promote vitamin D3 synthesis in humanskin. J Clin Endocrinol Metab. 1988; 67(2):373–8. PMID: 2839537

4. Becklund BR, Severson KS, Vang SV, DeLuca HF. UV radiation suppresses experimental autoimmuneencephalomyelitis independent of vitamin D production. PNAS. 2010; 107(14):6418–23. doi: 10.1073/pnas.1001119107 PMID: 20308557

5. Wang Y, Marling SJ, McKnight SM, Danielson AL, Severson KS, Deluca HF. Suppression of experi-mental autoimmune encephalomyelitis by 300–315nm ultraviolet light. Arch Biochem Biophys. 2013;536(1):81–6. doi: 10.1016/j.abb.2013.05.010 PMID: 23747577

6. Lucas R, Ponsonby A-L, Dear K, Valery P, Pender M, Taylor B, et al. Sun exposure and vitamin D areindependent risk factors for CNS demyelination. Neurology. 2011; 76(6):540–8. doi: 10.1212/WNL.0b013e31820af93d PMID: 21300969

7. Holick MF. Sunlight, UV-radiation, vitamin D and skin cancer: howmuch sunlight do we need? Sunlight,Vitamin D and Skin Cancer: Springer; 2008. p. 1–15.

8. Hilger J, Friedel A, Herr R, Rausch T, Roos F, Wahl DA, et al. A systematic review of vitamin D status inpopulations worldwide. Br J Nutr. 2014; 111(1):23–45. doi: 10.1017/S0007114513001840 PMID:23930771

9. Holick MF. Sunlight and vitamin D for bone health and prevention of autoimmune diseases, cancers,and cardiovascular disease. Am J Clin Nutr. 2004; 80(6 Suppl):1678s–88s. PMID: 15585788

10. Grant WB, Holick MF. Benefits and requirements of vitamin D for optimal health: a review. Altern MedRev. 2005; 10(2):94–111. PMID: 15989379

11. Orton SM, Wald L, Confavreux C, Vukusic S, Krohn JP, Ramagopalan SV, et al. Association of UV radi-ation with multiple sclerosis prevalence and sex ratio in France. Neurology. 2011; 76(5):425–31. doi:10.1212/WNL.0b013e31820a0a9f PMID: 21282589

12. Sloka JS, Pryse-Phillips WE, Stefanelli M. The relation of ultraviolet radiation and multiple sclerosis inNewfoundland. Can J Neurol Sci. 2008; 35(1):69–74. PMID: 18380280

13. van der Mei IA, Ponsonby AL, Blizzard L, Dwyer T. Regional variation in multiple sclerosis prevalencein Australia and its association with ambient ultraviolet radiation. Neuroepidemiology. 2001; 20(3):168–74. PMID: 11490162

14. Ascherio A, Munger KL, White R, et al. Vitamin D as an early predictor of multiple sclerosis activity andprogression. JAMANeurology. 2014; 71(3):306–14. doi: 10.1001/jamaneurol.2013.5993 PMID:24445558

15. Munger KL, Levin LI, Hollis BW, Howard NS, Ascherio A. Serum 25-hydroxyvitamin D levels and risk ofmultiple sclerosis. JAMA. 2006; 296(23):2832–8. PMID: 17179460

16. Mokry L, Ross S, Ahmad O, Forgetta V, Davey Smith G, Leong A, et al. Vitamin D and Risk of MultipleSclerosis: A Mendelian Randomization Study. PLoS Med. 2015; 12(8):e1001866. doi: 10.1371/journal.pmed.1001866 PMID: 26305103

17. Ramagopalan SV, Maugeri NJ, Handunnetthi L, Lincoln MR, Orton SM, Dyment DA, et al. Expressionof the multiple sclerosis-associated MHC class II Allele HLA-DRB1*1501 is regulated by vitamin D.PLoS Genet. 2009; 5(2):e1000369. doi: 10.1371/journal.pgen.1000369 PMID: 19197344

18. Ramagopalan SV, Dyment DA, Cader MZ, Morrison KM, Disanto G, Morahan JM, et al. Rare variantsin the CYP27B1 gene are associated with multiple sclerosis. Annals of Neurology. 2011; 70(6):881–6.doi: 10.1002/ana.22678 PMID: 22190362

19. Yu S, Cantorna MT. Epigenetic reduction in iNKT cells following in utero vitamin D deficiency in mice.Journal of immunology (Baltimore, Md: 1950). 2011; 186(3):1384–90.

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 16 / 18

20. van der Vliet HJ, von Blomberg BM, Nishi N, Reijm M, Voskuyl AE, van Bodegraven AA, et al. Circulat-ing V(alpha24+) Vbeta11+ NKT cell numbers are decreased in a wide variety of diseases that are char-acterized by autoreactive tissue damage. Clin Immunol. 2001; 100(2):144–8. PMID: 11465942

21. Dobson R, Giovannoni G, Ramagopalan S. The month of birth effect in multiple sclerosis: systematicreview, meta-analysis and effect of latitude. J Neurol Neurosurg Psychiatry. 2012:jnnp-2012-303934.

22. Webb AR, Engelsen O. Calculated ultraviolet exposure levels for a healthy vitamin D status. Photo-chem Photobiol. 2006; 82(6):1697–703. PMID: 16958558

23. Rhodes LE, Webb AR, Fraser HI, Kift R, Durkin MT, Allan D, et al. Recommended Summer SunlightExposure Levels Can Produce Sufficient (&ges; 20 ng ml− 1) but Not the Proposed Optimal (&ges; 32ng ml− 1) 25 (OH) D Levels at UK Latitudes. J Invest Dermatol. 2010; 130(5):1411–8. doi: 10.1038/jid.2009.417 PMID: 20072137

24. Mavroeidi A, Aucott L, Black AJ, Fraser WD, Reid DM, Macdonald HM. Seasonal Variation in 25 (OH)D at Aberdeen (57° N) and Bone Health Indicators–Could Holidays in the Sun and Cod Liver Oil Sup-plements Alleviate Deficiency? 2013.

25. Visser EM, Wilde K, Wilson JF, Yong KK, Counsell CE. A new prevalence study of multiple sclerosis inOrkney, Shetland and Aberdeen city. J Neurol Neurosurg Psychiatry. 2012; 83(7):719–24. doi: 10.1136/jnnp-2011-301546 PMID: 22577232

26. McQuillan R, Leutenegger A- L, Abdel-Rahman R, Franklin CS, Pericic M, Barac-Lauc L, et al. Runs ofhomozygosity in European populations. The American Journal of Human Genetics. 2008; 83(3):359–72. doi: 10.1016/j.ajhg.2008.08.007 PMID: 18760389

27. Tenesa A, Farrington SM, Prendergast JG, Porteous ME,Walker M, Haq N, et al. Genome-wide associ-ation scan identifies a colorectal cancer susceptibility locus on 11q23 and replicates risk loci at 8q24and 18q21. Nat Genet. 2008; 40(5):631–7. doi: 10.1038/ng.133 PMID: 18372901

28. Seamans KM, Cashman KD. Existing and potentially novel functional markers of vitamin D status: asystematic review. The American journal of clinical nutrition. 2009:ajcn. 27230D.

29. Wallace A, Gibson S, De La Hunty A, Lamberg-Allardt C, Ashwell M. Measurement of 25-hydroxyvita-min D in the clinical laboratory: current procedures, performance characteristics and limitations. Ste-roids. 2010; 75(7):477–88. doi: 10.1016/j.steroids.2010.02.012 PMID: 20188118

30. Ross AC, Manson JE, Abrams SA, Aloia JF, Brannon PM, Clinton SK, et al. The 2011 report on dietaryreference intakes for calcium and vitamin D from the Institute of Medicine: what clinicians need to know.The Journal of Clinical Endocrinology & Metabolism. 2011; 96(1):53–8.

31. Pearce SH, Cheetham TD. Diagnosis and management of vitamin D deficiency. BMJ. 2010; 340:b5664. doi: 10.1136/bmj.b5664 PMID: 20064851

32. Zgaga L, Theodoratou E, Farrington SM, Agakov F, Tenesa A, Walker M, et al. Diet, environmental fac-tors, and lifestyle underlie the high prevalence of vitamin D deficiency in healthy adults in Scotland, andsupplementation reduces the proportion that are severely deficient. J Nutr. 2011; 141(8):1535–42. doi:10.3945/jn.111.140012 PMID: 21697298

33. SIMD. Scottish Parliamentary Constituency Profile—00410739.pdf. In: Deprivation SIoM, editor.Scotland2012.

34. Vyas S, Kumaranayake L. Constructing socio-economic status indices: how to use principal compo-nents analysis. Health Policy Plan. 2006; 21(6):459–68. PMID: 17030551

35. Nakao K, Treas J. Computing 1989 occupational prestige scores: publisher not identified; 1990.

36. van Buuren S, Goothuis-Oudshoorn K. mice: Multivariate Imputation by Chained Equations in R. Jour-nal of Statistical Software. 2011; 45(3):1–67.

37. Team RC. R: A language and environment for statistical computing. R Foundation for Statistical Com-puting, Vienna, Austria. URL http://www.R-project.org/. 2015.

38. Holick MF, Binkley NC, Bischoff-Ferrari HA, Gordon CM, Hanley DA, Heaney RP, et al. Evaluation,treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline. TheJournal of Clinical Endocrinology & Metabolism. 2011; 96(7):1911–30.

39. MacLaughlin J, Holick MF. Aging decreases the capacity of human skin to produce vitamin D3. J ClinInvest. 1985; 76(4):1536. PMID: 2997282

40. Mithal A, Wahl D, Bonjour J-P, Burckhardt P, Dawson-Hughes B, Eisman J, et al. Global vitamin D sta-tus and determinants of hypovitaminosis D. Osteoporos Int. 2009; 20(11):1807–20. doi: 10.1007/s00198-009-0954-6 PMID: 19543765

41. Brock K, HuangW-Y, Fraser D, Ke L, Tseng M, Stolzenberg-Solomon R, et al. Low vitamin D status isassociated with physical inactivity, obesity and low vitamin D intake in a large US sample of healthymiddle-aged men and women. The Journal of steroid biochemistry and molecular biology. 2010; 121(1):462–6.

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 17 / 18

42. Jorde R, Sneve M, Emaus N, Figenschau Y, Grimnes G. Cross-sectional and longitudinal relationbetween serum 25-hydroxyvitamin D and body mass index: the Tromsø study. Eur J Nutr. 2010; 49(7):401–7. doi: 10.1007/s00394-010-0098-7 PMID: 20204652

43. Wortsman J, Matsuoka LY, Chen TC, Lu Z, Holick MF. Decreased bioavailability of vitamin D in obesity.The American journal of clinical nutrition. 2000; 72(3):690–3. PMID: 10966885

44. Wellens RI, Roche AF, Khamis HJ, Jackson AS, Pollock ML, Siervogel RM. Relationships between thebody mass index and body composition. Obes Res. 1996; 4(1):35–44. PMID: 8787936

Vitamin D in Orkney

PLOS ONE | DOI:10.1371/journal.pone.0155633 May 17, 2016 18 / 18